MF.07 Measurement of outcome in cognitive behavioural rehabilitation services: HoNOS-ABI & FIMFAM-UK sensitivity to change and lessons from HoNOS Rasch analysis. Issue 8 (7th July 2011)

• Record Type: Journal Article
• Authors: Bajo, A; Potter, S; Bateman, A; Lazaro, F; Fleminger, S
• Is Part Of: Journal of neurology, neurosurgery and psychiatry. Volume 82:Issue 8(2011)
• Page Start: e2
• Page End: e2
• Publication Date: 2011-07-07
• Subjects: Neurology -- Periodicals; Nervous system -- Surgery -- Periodicals; Psychiatry -- Periodicals; 616.8
• DOI: 10.1136/jnnp-2011-300504.18
• Languages: English
• ISSNs: 0022-3050

Abstract

Abstract : Aims: To determine HoNOS-ABI and FIMFAM-UK sensitivity to change following cognitive-behavioural rehabilitation for brain injury, and to examine the psychometric properties of these outcome measures. Methods: 324 brain injured survivors underwent cognitive-behavioural rehabilitation at one of three participating centres: Lishman, Blackheath, and Edgware. Effect sizes of pre- and post-rehabilitation ratings were examined at each centre. 194 HoNOS-ABI pre-rehabilitation ratings collected across all centres were entered into an analysis using RUMM2020 (Tennant et al ). Results: Preliminary results for sensitivity to change and Rasch analyses are reported. Sensitivity: Moderate to large effect sizes were found at individual centres for HoNOS and FAM. HoNOS was the most sensitive tool at Blackheath and LBIU, and FAM was the most sensitive measure at Edgware. Rasch: HoNOS-ABI was found to show significant misfit to the Rasch model. All but one of the questions (cognitive) showed disordered thresholds. A person separation index of 0.63 suggested the HoNOS to be internally unreliable. Visual inspection of Category Probability Curves was used to guide rescoring. A Principal Components Analysis on the Rasch-transformed data was used to identify two possible subscales within HoNOS-ABI with potentially better psychometric properties. Conclusions: HoNOS-ABI and FAM total scores were moderately to highly sensitive to change following cognitive-behavioural rehabilitation. HoNOS was better able to pick up change at Blackheath and LBIU, whereas FAM would be the measure of choice at Edgware. Rasch analysis proved to be informative of the strengths and weaknesses of HoNOS as a neuropsychiatric outcome measure, and highlighted that further work is needed to refine it as a tool for rehabilitation outcome measurement. HoNOS total score showed sensitivity to change. However, psychometric issues highlighted with this analysis indicated that a modified version may make this tool more psychometrically robust and may even increase its sensitivity to change.